A hallmark of human development is the ability to make adaptive changes in perception that occurred as a result of experience. The internal representations that influence perception must be constructed from multiple levels of statistical information that are present in the natural environment. To illustrate what is meant by statistical information, after hearing a bark, there is a higher probability of seeing a dog than seeing a cat (i.e. p(Vdog | Adog) > p(Vcat | Adog). While it is widely believed that information acquired via statistical learning is an essential component of experience-based perceptual development, little is known about the neurobiological and behavioral mechanisms that translate experience into lasting perceptual change or how variations in these mechanisms can lead to atypical or delayed development. This proposal addresses three major gaps in our knowledge of the mechanisms of experience-based perceptual development: 1) the effect of statistical information on neural activity in sensory cortices in infancy; 2) whether deviations in these neural responses predict developmental delays in a high-risk population; 3) the relationship between statistically-induced changes in neural activity and behavioral changes in perception. Aim 1 Characterize the effects of statistical information on neural activity early in postnatal development. Typically developing, 6-month-olds will undergo near infrared spectroscopy (NIRS) recording to reveal changes in neural activity in auditory and visual sensory cortices during and after experience with cross-modal statistical information. We hypothesize that infant sensory cortex will exhibit significant changes in activity (i.e., increased activity during an unexpected visual omission) as result of statistical experience. Aim 2 Examine deviations in statistically-induced neural activity in at-risk populations. If statistically-induced neural responses support typical development in numerous domains, abnormal neural responses to statistical information will predict poor behavioral outcomes later in postnatal life and establish biomarkers that could assist in ameliorating abnormal development. Using data from Aim 1 as a comparison group, we will characterize neural responses in infants at high risk for delays in perceptual and cognitive develop- ment as a result of being born extremely preterm (< 28 weeks). Aim 3 Determine the relationship between statistically-induced neural activity and perceptual change. Visual cortex activity after a visual stimulus produces a visual percept, but the behavioral consequences of visual cortex activity during the unexpected omission of a stimulus are unknown yet are essential to understanding how statistically-induced neural changes support perceptual development. Guided by specific hypotheses, we will correlate neural responses to unexpected omissions to behavior on perceptual tasks.